Closed TotalDay closed 1 year ago
Hi, we haven't added support for SD2.1 yet. But you can checkout this notebook for SD2.1 model conversion to openvino format - https://github.com/openvinotoolkit/openvino_notebooks/blob/main/notebooks/236-stable-diffusion-v2/236-stable-diffusion-v2-text-to-image.ipynb
Thanks.
how to add more item to this list?
like dreamlike-art/dreamlike-photoreal-2.0 and many other
Where i can change width and height? i want 768 x 768 \ 768 x 576 \ 1024 x 576
In order to add more models to the list, you will need to modify the code. Once you have openvino version of the models, you can search for all the occurrences of "SD_1.5" in the code base and follow similar method to add custom models (based on 1.4 or 1.5)
Regarding the model dreamlike-art/dreamlike-photoreal-2.0, I believe you can only use it free on https://dreamlike.art/.
Currently we have only added support for SD-1.4 and SD-1.5 both of which are trained on 512x512 images and produces 512x512 images.
I converted dreamlike-art/dreamlike-photoreal-2.0 - ant it work!!!! Today i converted more 10 models from https://huggingface.co/models?pipeline_tag=text-to-image&sort=downloads
at Intel ARC 750 speed - 7.26 it/sec!!!! More than RTX 3060
Thanks for the update @TotalDay -- so you were able to integrate those models into your local running version of ov-stable-diffusion GIMP plugin?
Thanks for the update @TotalDay -- so you were able to integrate those models into your local running version of ov-stable-diffusion GIMP plugin?
yes, all models from hugginface.co work i am using converter in the Jupiter Notebook (local)
Work on My Intel ARC 750 8GB 7.2it/sec
Super Thanks!
How can i use Stable Diffusion 2 to openvino format?